PyCon debrief #2: The talks you (and I) missed
There was some complaints about the quality of the talks this year. I don’t know, I think they were at least as good as the ones at EuroPython and the Plone conference. Were they really better the years before? Or is it just that the people who complain has become more advanced, and that they expect the talks to become more advanced as well?
I did see some really interesting talks. Not all of the speakers are great speakers, but the content of the following talks were very interesting. The first talk I saw was:
Repoze is a project to basically replace the monolithic Zope publishers with a layered WSGI-type hierarchy of middle ware. It’s progressing rapidly and is really cool. Repoze turned out to become the PyCon main theme for me, and PyCon was tied together when Repoze returned with the last talk I saw:
Mark explained a bit of the philosophy behind TurboGears, and why TurboGears decided to ditch SQLObject and CherryPy and switched to SQLAlchemy and Pylons instead. It’s all about using the best technology available and not being a monolithic dictator. Pretty much, not making one of the major mistakes of Zope 2, which I took up in my talk. Repoze comes into this because Mark Ramm and the other TG people are just as excited by repoze as the zope-guys are, and are gearing up to use some of the repoze components.
Adrian Holovaty: The State of Django and
James Bennett: Developing reusable Django applications
was also interesting, for exactly the opposite reason. More on this, repoze and web framweorks is probably gonna come in another blog post from me.
Matt talked about how to measure the complexity of your code and how this relates to how much testing you need to do, and why straight line coverage measurements is not a good measurement but better than none).
Things I missed but people told me was very good was:
Brett Cannon: How Import Does Its Thing and
Raymond D Hettinger: Core Python Containers — Under the Hood
Python under the hood talks seem more interesting than I thought that they were. I should have seen these.
Brandon C Rhodes: Using Grok to Walk Like a Duck
I already use the Zope component architecture, so I don’t need convincing. However, Brandons slides do a good job in explaining why it’s a good thing to use.
This also got good words from others. Evidently Trac is much more flexible and easier to extend than people imagine. Or Noah is just super-pedagogical. 🙂
All in all a good share of interesting talks, in my opinion.